Probability-based task scheduling method for minimizing energy consumption in embedded system

Abstract

The invention provides a probability-based task scheduling method for minimizing energy consumption in an embedded system and is applied to an embedded heterogeneous multiprocessor system. The method comprises the following steps of firstly initializing an ASAP (as soon as possible) or ALAP (as later as possible) algorithm to generate CPU (central processing unit) allocation schemes satisfying a data dependence relationship, then, specifying an execution voltage onto a CPU allocated to each task by using a dynamic planning method, and selecting an allocation scheme of minimum energy consumption from the CPU allocation schemes. The method also comprises the following steps of generating disturbance based on a current optimum allocation scheme to generate a set of new CPU allocation schemes satisfying a task dependence relationship, then, allocating a voltage to each CPU allocation scheme by adopting the dynamic planning method, and selecting a final CPU and voltage allocation scheme of the minimum energy consumption. The invention is capable of providing a scheduling scheme which is more effective and energy consumption-saving, is capable of satisfying the time limit of a real-time system by specified probability, and not only is suitable for task scheduling in a soft real-time system but also is suitable for the task scheduling in a hard real-time system.

Description

technical field [0001] The invention belongs to the embedded field, and in particular relates to a task scheduling method for minimizing energy consumption in a probability-based embedded heterogeneous multiprocessor system. Background technique [0002] In an embedded system that uses batteries as the main energy source, it is of great significance to save energy consumption as much as possible while ensuring system performance to prolong the service life of the embedded system. Dynamic voltage scaling (DVS) is considered to be a more effective technology in low-power design. DVS technology is mainly based on the fact that the energy consumption of the processor is proportional to the square of the operating voltage. Real-time adjustment of processor operating voltage can well solve the performance and power consumption requirements in embedded real-time systems. [0003] In embedded systems, the execution time of tasks is often affected by uncertain environmental factors ...

AI Technical Summary

Problems solved by technology

However, in actual situations, the processor scheduling scheme of tasks is often unknown, so this scheduling strategy is not applicable in actual situations

Images